Top 10 Lessons Learned
In our ongoing shift of Europeana as a portal to Europeana as a
platform
David Haskiya, Product Development Manager
Europeana is an organisation that is
passionate about cultural innovation.
Sitting at the intersection of culture and
technology, our aim is to open up Europe's
creative and cultural wealth to as wide an
audience as possible.
Because Cultural Innovation:
- Advances society both economically and socially 
- Unites Europe through a shared cultural heritage 
- Celebrates the wealth of cultures across Europe 
- Enables personal growth and development 
My Top 10 Lessons Learned
Lesson #1: APIs are means and not
ends
It's not just about the technology

It's what you as an organisation and business can
accomplish with an API that you can't without

An API needs to be carefully packaged into a number of
other activities in order to fully succeed

Since I'm not a technologist I won't be sharing any lessons
on technical API-design, but on the lessons I've drawn
from the mistakes and successes we've made at
Europeana over the last 3-4 years as APIs have gone from
a loose idea towards a core service
Lesson #2: Show, don't tell...
...is key to getting buy-in

In my organisation very few colleagues and even fewer
partners knew what an API was in 2010

A lot of effort was needed to get full organisational buy-in

Getting buy-in from our funders was difficult. You need to
be able to explain in it in terms of business that they
understand
•
Having a couple of applications helps a lot

The buzz of the hackathons we arranged in 2011 were key
in getting internal and external buy-in

You need to be able to measure success. Quantitatively,
qualitatively, convincingly.
•
Yes, non-profit organisation or government agency, that
means you too!
Lesson #3: Copyright is the killer of
confidence
Eliminate doubt and confusion!

Usually the most obscure part of your Terms of Use and
thus a main cause of non-adoption

Can stop you cold even if you've done all the technology
aspects right
•
Our API was semi-open for too long! It caused bad-will and
doubt among developers

And is usually much more difficult and slower to change
than what your technology
•
It took us almost 2 years to convince all of our data
providing partners to CC0 their metadata and adopt a
structured licensing framework for media served

Adopt standard licences as much as possible and support
the public domain
Lesson #4: When the guerilla war is
won, governing begins
From guerilla to governor

In many organisations APIs are developed guerilla style at
first by one or two firebrands, skunkworks style.

When you get organisational buy-in APIs become truly part
of your business with all that it entails

The guerilla phase tends to be more fun. Your explore new
concepts and technologies and you are the consensus

When the API becomes part of the line business the
meetings set in, the overhead, the consensus building...

Now you need to govern. It's not for everyone. Either move
on or soldier on. And if you move on, don't kibbitz your
successor.
Lesson #5: Document your data
Data usability is essential

Content is as key to API-design as it is to web design or
app development
• Document it with as much care as you do for your API, if
not more
•
We learned this lesson late! Now we're investing in featuring
the best of our data

Investing in data quality always pays off in the long-run,
whereas code has a short half-life
•
Try to solve fundamental data issues at source, rather than
hide them by code
•
When aggregating data from multiple sources align to one
model and normalize strictly!
Lesson #6: Build it and they will NOT
come
Your API is here! How do you tell the
world?

Unless you're extraordinarily lucky you need to to market
and communicate your API for it to get traction with
developers and customers

The Zen paradox of API-marketing:
•
An API is like any other product and must be marketed like
any other product in order to be succesful
•
An API is not like other products and cannot be succesfully
marketed like any other products

This is probably a fake paradox. But it can be difficult for
an organisation get the perspective and balance right.
•
We're still working on it
Lesson #7: Developers are users too!
They're not THAT special

Just because they're power-users of the web doesn't mean
you shouldn't invest in API-documentation or developer
portal UX

The methods and processes of UX-design can be
succesfully adapted and applied to APIs and developer
portals
•
Our soon to be released developer portal is our first we're
we've researched and designed as we would a consumer
site
•
We're still not as good at documentation as I would want us
to be
Lesson #8: Hackathons ≠ Developer
Outreach
The gospel of API-evangelism

Consider what hackathons are good for and not
•
In my sector (libraries, archives, museums) I often feel they're
held because it's what everyone does

Fold your hackathons into along-term outreach/community
programme
•
Community outreach is labour intensive one marketeer doing it
with 10% of of his or her attention won't cut it

Be inclusive when you organise them and don't profiteer
on the efforts of the participating developers
•
There's an increasing body of hackathon best practices
Lesson #9: Some lessons you need to
learn yourself. Again. And again.
You will #FAIL

If you're new to developing APIs and the business of APIs
you will do your research carefully and make mistakes
anyway
•
I remember listing many pitfalls to my management when we
started with our API development. Then I fell in them.
•
I think the business related pitfalls are deeper and easier to fall
into but that could be because I'm not a developer.
•
Quick iterations makes for shallow pits discovered early

If you're experienced in developing APIs and running a
API-centric business you will make mistakes anyway
•
You can hopefully climb out the pit faster the second time

The government sector can have a very hard time
accepting that you can fail forward
Lesson #10: I actually don't have one,
but I'm willing to listen
That was it! Questions? Comments?
Grab me in a break or get in touch.
Email: david.haskiya@kb.nl
Twitter: @david.haskiya

Top 10 Lessons Learned - In our ongoing shift from portal to platform

  • 1.
    Top 10 LessonsLearned In our ongoing shift of Europeana as a portal to Europeana as a platform David Haskiya, Product Development Manager
  • 2.
    Europeana is anorganisation that is passionate about cultural innovation. Sitting at the intersection of culture and technology, our aim is to open up Europe's creative and cultural wealth to as wide an audience as possible. Because Cultural Innovation: - Advances society both economically and socially  - Unites Europe through a shared cultural heritage  - Celebrates the wealth of cultures across Europe  - Enables personal growth and development 
  • 3.
    My Top 10Lessons Learned
  • 4.
    Lesson #1: APIsare means and not ends
  • 5.
    It's not justabout the technology  It's what you as an organisation and business can accomplish with an API that you can't without  An API needs to be carefully packaged into a number of other activities in order to fully succeed  Since I'm not a technologist I won't be sharing any lessons on technical API-design, but on the lessons I've drawn from the mistakes and successes we've made at Europeana over the last 3-4 years as APIs have gone from a loose idea towards a core service
  • 6.
    Lesson #2: Show,don't tell...
  • 7.
    ...is key togetting buy-in  In my organisation very few colleagues and even fewer partners knew what an API was in 2010  A lot of effort was needed to get full organisational buy-in  Getting buy-in from our funders was difficult. You need to be able to explain in it in terms of business that they understand • Having a couple of applications helps a lot  The buzz of the hackathons we arranged in 2011 were key in getting internal and external buy-in  You need to be able to measure success. Quantitatively, qualitatively, convincingly. • Yes, non-profit organisation or government agency, that means you too!
  • 8.
    Lesson #3: Copyrightis the killer of confidence
  • 9.
    Eliminate doubt andconfusion!  Usually the most obscure part of your Terms of Use and thus a main cause of non-adoption  Can stop you cold even if you've done all the technology aspects right • Our API was semi-open for too long! It caused bad-will and doubt among developers  And is usually much more difficult and slower to change than what your technology • It took us almost 2 years to convince all of our data providing partners to CC0 their metadata and adopt a structured licensing framework for media served  Adopt standard licences as much as possible and support the public domain
  • 10.
    Lesson #4: Whenthe guerilla war is won, governing begins
  • 11.
    From guerilla togovernor  In many organisations APIs are developed guerilla style at first by one or two firebrands, skunkworks style.  When you get organisational buy-in APIs become truly part of your business with all that it entails  The guerilla phase tends to be more fun. Your explore new concepts and technologies and you are the consensus  When the API becomes part of the line business the meetings set in, the overhead, the consensus building...  Now you need to govern. It's not for everyone. Either move on or soldier on. And if you move on, don't kibbitz your successor.
  • 12.
  • 13.
    Data usability isessential  Content is as key to API-design as it is to web design or app development • Document it with as much care as you do for your API, if not more • We learned this lesson late! Now we're investing in featuring the best of our data  Investing in data quality always pays off in the long-run, whereas code has a short half-life • Try to solve fundamental data issues at source, rather than hide them by code • When aggregating data from multiple sources align to one model and normalize strictly!
  • 14.
    Lesson #6: Buildit and they will NOT come
  • 15.
    Your API ishere! How do you tell the world?  Unless you're extraordinarily lucky you need to to market and communicate your API for it to get traction with developers and customers  The Zen paradox of API-marketing: • An API is like any other product and must be marketed like any other product in order to be succesful • An API is not like other products and cannot be succesfully marketed like any other products  This is probably a fake paradox. But it can be difficult for an organisation get the perspective and balance right. • We're still working on it
  • 16.
    Lesson #7: Developersare users too!
  • 17.
    They're not THATspecial  Just because they're power-users of the web doesn't mean you shouldn't invest in API-documentation or developer portal UX  The methods and processes of UX-design can be succesfully adapted and applied to APIs and developer portals • Our soon to be released developer portal is our first we're we've researched and designed as we would a consumer site • We're still not as good at documentation as I would want us to be
  • 18.
    Lesson #8: Hackathons≠ Developer Outreach
  • 19.
    The gospel ofAPI-evangelism  Consider what hackathons are good for and not • In my sector (libraries, archives, museums) I often feel they're held because it's what everyone does  Fold your hackathons into along-term outreach/community programme • Community outreach is labour intensive one marketeer doing it with 10% of of his or her attention won't cut it  Be inclusive when you organise them and don't profiteer on the efforts of the participating developers • There's an increasing body of hackathon best practices
  • 20.
    Lesson #9: Somelessons you need to learn yourself. Again. And again.
  • 21.
    You will #FAIL  Ifyou're new to developing APIs and the business of APIs you will do your research carefully and make mistakes anyway • I remember listing many pitfalls to my management when we started with our API development. Then I fell in them. • I think the business related pitfalls are deeper and easier to fall into but that could be because I'm not a developer. • Quick iterations makes for shallow pits discovered early  If you're experienced in developing APIs and running a API-centric business you will make mistakes anyway • You can hopefully climb out the pit faster the second time  The government sector can have a very hard time accepting that you can fail forward
  • 22.
    Lesson #10: Iactually don't have one, but I'm willing to listen
  • 23.
    That was it!Questions? Comments? Grab me in a break or get in touch. Email: david.haskiya@kb.nl Twitter: @david.haskiya